Rapamycin (sirolimus) (FIG. 1) is a lipophilic macrolide produced by Streptomyces hygroscopicus NRRL 5491 (Sehgal et al., 1975; Vézina et al., 1975; U.S. Pat. No. 3,929,992; U.S. Pat. No. 3,993,749) with a 1,2,3-tricarbonyl moiety linked to a pipecolic acid lactone (Paiva et al., 1991). Other related macrolides (FIG. 2) include FK506 (tacrolimus) (Schreiber and Crabtree, 1992), FK520 (ascomycin or immunomycin) (Wu et al., 2000), FK525 (Hatanaka H, et al., 1989, FK523 (Hatanaka, H., et al., 1988), antascomicins (Fehr, T., et al., 1996) and meridamycin (Salituro et al., 1995). For the purpose of this invention rapamycin is described by the numbering convention of McAlpine et al. (1991) in preference to the numbering conventions of Findlay et al. (1980) or Chemical Abstracts (11th Cumulative Index, 1982-1986 p60719CS).
The versatile mode of action of rapamycin demonstrates the pharmacological value of the compound and emphasizes the necessity to isolate novel derivatives of the drug. Rapamycin shows moderate antifungal activity, mainly against Candida species but also against filamentous fungi (Baker et al., 1978; Sehgal et al., 1975; Vézina et al., 1975; U.S. Pat. No. 3,929,992; U.S. Pat. No. 3,993,749). Rapamycin inhibits cell proliferation by targeting signal transduction pathways in a variety of cell types, e.g. by inhibiting signalling pathways that allow progression from the G1 to the S-phase of the cell cycle (Kuo et al., 1992). In T cells rapamycin inhibits signalling from the IL-2 receptor and subsequent autoproliferation of the T cells resulting in immunosuppression. The inhibitory effects of rapamycin are not limited to T cells, since rapamycin inhibits the proliferation of many mammalian cell types (Brunn et al, 1996). Rapamycin is, therefore, a potent immunosuppressant with established or predicted therapeutic applications in the prevention of organ allograft rejection and in the treatment of autoimmune diseases (Kahan et at, 1991). It appears to cause fewer side effects than the standard anti-rejection treatments (Navia, 1996). 40-O-(2-hydroxy)ethyl-rapamycin (SDZ RAD, Certican, Everolimus) is a semi-synthetic analogue of rapamycin that shows immunosuppressive pharmacological effects (Sedrani, R. et at, 1998; U.S. Pat. No. 5,665,772). The clinical efficacy of the drug is presently under investigation in Phase III clinical trials (Kirchner et at, 2000). The rapamycin ester CCI-779 (Wyeth-Ayerst) inhibits cell growth in vitro and inhibits tumour growth in vivo (Yu et at, 2001). The drug is currently in Phase III clinical trials. The value of rapamycin in the treatment of chronic plaque psoriasis (Kirby and Griffins, 2001), the potential use of effects such as the stimulation of neurite outgrowth in PC12 cells (Lyons et at, 1994), the block of the proliferative responses to cytokines by vascular and smooth muscle cells after mechanical injury (Gregory et al., 1993) and its role in prevention of allograft fibrosis (Waller and Nicholson, 2001) are areas of intense research (Kahan and Camardo, 2001). Recent reports reveal that rapamycin is associated with lower incidence of cancer in organ allograft patients on long-term immunosuppressive therapy than those on other immunosuppressive regimes, and that this reduced cancer incidence is due to inhibition of angiogenesis (Guba et at, 2002). It has been reported that the neurotrophic activities of immunophilin ligands are independent of their immunosuppressive activity (Steiner et al., 1997) and that nerve growth stimulation is promoted by disruption of the mature steroid receptor complex as outlined in the patent application WO01/03692. Side effects such as hyperlipidemia and thrombocytopenia as well as potential teratogenic effects have been reported (Hentges et at, 2001; Kahan and Camardo, 2001).
The polyketide backbone of rapamycin is synthesised by head-to-tail condensation of a total of seven propionate and seven acetate units to a shikimate derived cyclohexane carboxylic acid starter unit (Paiva et al., 1991). The L-lysine derived imino acid, pipecolic acid, is condensed via an amide linkage onto the last acetate of the polyketide backbone (Paiva et al., 1993) and is followed by lactonisation to form the macrocycle. A 107 kb genomic region containing the biosynthetic gene cluster has been sequenced (Schwecke et at, 1995). Analysis of the open reading frames revealed three large genes encoding the modular polyketide synthase (PKS) (Aparicio et al., 1996; Schwecke et al., 1995). Embedded between the PKS genes lies the rapP gene encoding a protein with sequence similarity to activation domains of nonribosomal peptide synthetases and it is thought to act analogously (König et al., 1997). The region encoding the PKS genes is flanked on both sides by 24 additional open reading frames encoding enzymes believed to be required for the biosynthesis of rapamycin (Molnár et al., 1996). These include the following post-polyketide modification enzymes: two cytochrome P450 monooxygenases, designated as RapJ and RapN, an associated ferredoxin RapO, and three potential SAM-dependent O-methyltransferases RapI, RapM and RapQ. Other adjacent genes have putative roles in the regulation and the export of rapamycin (Molnár et al., 1996). The cluster also contains the gene rapL whose product RapL is proposed to catalyse the formation of the rapamycin precursor L-pipecolic acid through the cyclodeamination of L-lysine (Khaw et al., 1998; Paiva et al., 1993). The introduction of a frameshift mutation into rapL gave rise to a mutant unable to produce significant amounts of rapamycin and feeding of L-pipecolic acid to the growth medium restored wild-type levels of rapamycin production (Khaw et al., 1998). The biosynthetic precursors to the cyclohexane ring of rapamycin originate from the shikimic acid pathway (Lowden et al., 1996; Lowden et al., 2001). Other closely-related macrolides such as FK506 (tacrolimus) (Schreiber and Crabtree, 1992), FK520 (ascomycin or immunomycin) (Wu et al., 2000), antascomicin (Fehr, T., et al., 1996) and meridamycin (Salituro et al., 1995) share a common pharmacophore that interacts with FK506-binding proteins (FKBPs) (FIG. 2). Thus rapamycin and related compounds for example, but without limitation, FK506, FK520, ‘hyg’, FK523, meridamycin, antascomicin, FK525 and tsukubamydn can be considered “FKBP-ligands”. The partial sequence of the FK506 gene cluster (Motamedi et al., 1996; Motamedi et al., 1997; Motamedi and Shafiee, 1998), the ‘hyg’ cluster (Ruan et al., 1997) and the complete sequence of the FK520 gene cluster have been published (VWu et al., 2000; U.S. Pat. No. 6,150,513). There is significant homology between genes within these clusters and the rapamycin biosynthetic gene cluster and similarity in enzyme function (Motamedi et al., 1996).
The pharmacologic actions of rapamycin characterised to date are believed to be mediated by the interaction with cytosolic receptors termed FKBPs or immunophilins. Immunophilins (this term is used to denote immunosuppressant binding proteins) catalyse the isomerisation of cis and trans peptidyl-proline bonds and belong to a highly conserved family of enzymes found in a wide variety of organisms (Rosen and Schreiber, 1992). Two large groups of enzymes belonging to the family of immunophilins are represented by FKBPs and cyclophilins (Schreiber and Crabtree, 1992). The major intracellular rapamycin receptor in eukaryotic T-cells is FKBP12 (DiLella and Craig, 1991) and the resulting complex interacts specifically with target proteins to inhibit the signal transduction cascade of the cell. FK506, an immunosuppressive agent structurally related to rapamycin, also specifically binds to FKBP12 but it effects immunosuppression through a different mechanism (Chang et al.,1991; Sigal and Dumont, 1992). Rapamycin and FK506 compete for the same binding site, thus FK506 can have an antagonistic effect with rapamycin when the two drugs are used together (Cao et al., 1995). Analysis of the crystal structure of the FKBP12-rapamycin complex has identified a rapamycin-binding pharmacophore termed the ‘binding domain’ (Van Duyne et al., 1993) (see FIG. 1). The ‘binding domain’ is required for the interaction with the immunophilin and consists, for both FK506 and rapamycin, of the C-1 to C-14 region including the ester linkage, the pipecolinyl ring, the dicarbonyl and the hemiketal ring (see FIG. 2). The interaction is characterised by many hydrophobic contacts and some hydrogen bonds including one to the hydroxyl group on the cyclohexane ring. The pipecolinyl ring (C2 to N7) makes the deepest penetration into the protein where it is surrounded by highly conserved aromatic amino acid residues lining the hydrophobic binding cavity. Both the C1 and the C8 carbonyl groups are involved in hydrogen bonding and the C9 carbonyl group protrudes into a pocket formed by three completely conserved aromatic amino acid residues (one tyrosine and two phenylalanine acid residues) in FKBP12. The domain of the immunophilin-ligand complex interacting with the target protein projects away from FKBP.
The target of the rapamycin-FKBP12 complex has been identified in yeast as TOR (target of rapamycin) (Alarcon et al., 1999) and the mammalian protein is known as FRAP (FKBP-rapamycin associated protein) or mTOR (mammalian target of rapamycin) (Brown et al., 1994). These proteins show significant similarity to the phosphotransferase domains of phosphatidylinositol 3-kinases and the observation that a point mutation in the FKBP12-rapamycin binding domain (FRB) of mTOR abolishes mTOR kinase activity provides evidence for the involvement of FRB in the function of the kinase domain (Vilella-Bach et al, 1999). The crystal structure of FKBP12-rapamycin with a truncated form of mTOR containing the FRB domain (Chen et al., 1995) has been obtained thus defining the ‘effector’ domain of rapamycin (Choi et al, 1996; Liang et al, 1999). The analysis of the crystal structure revealed that protein-protein contacts are relatively limited compared to the interaction between rapamycin and each protein. No hydrogen bonds between rapamycin and FRB were identified. Interaction is concentrated in a series of hydrophobic contacts between the triene region of rapamycin and mainly aromatic residues of FRB (Liang et al., 1999). The most deeply buried atom of rapamycin is the methyl attached to C23 (see FIG. 2). The C23 to C34 region and the cyclohexyl ring of rapamycin make superficial hydrophobic contacts with FRB. A small conformational change in rapamycin was evident between the binary and the ternary complexes (Liang et al., 1999).
Divergences between the biological effects of C16 methoxy group rapamycin analogues and their ability to bind FKBP12 were detected and the location of the C16 substituents at the interfacial space between FKBP12 and mTOR was postulated (Luengo et al., 1995). The analysis of the crystal structure of FKBP12 with the non-immunosuppressive 28-O-methyl rapamycin revealed a significant difference in the orientation of the cyclohexyl ring which may result in disruption of mTOR binding (Kallen et al., 1996).
Rapamycin impacts signalling cascades within the cell through the inhibition of the p70S6k kinase, a serine/threonine kinase in higher eukaryotes which phosphorylates the ribosomal protein S6 (Ferrari et al., 1993; Kuo et al., 1992). The S6 protein is located in the ribosomal 40S subunit and it is believed to be an important functional site involved in tRNA and mRNA binding. A regulatory function for mRNA translation through S6 phosphorylation by p70S6k has been postulated (Kawasome et al., 1998). Rapamycin inhibits protein synthesis through its effect on other growth related events, including the activity of cyclin-dependent kinases, phosphorylation of cAMP-responsive element modulator (CREM) and phosphorylation of the elongation factor binding protein 4E-BP-1 (PHAS1) (Hung et al., 1996). The drug induces the accumulation of the dephosphorylated species of 4E-BP1 that binds to the translation initiation factor eIF-4E, thus, suppressing translation initiation of cap-dependent mRNAs (Hara et al., 1997; Raught et al., 2001).
A link between mTOR signalling and localized protein synthesis in neurons; the effect on the phosphorylation state of proteins involved in translational control; the abundance of components of the translation machinery at the transcriptional and translational levels; control of amino acid permease activity and the coordination of the transcription of many enzymes involved in metabolic pathways have been described (Raught et al., 2001). Rapamycin sensitive signalling pathways also appear to play an important role in embryonic brain development, learning and memory formation (Tang et al, 2002). Research on TOR proteins in yeast also revealed their roles in modulating nutrient-sensitive signalling pathways (Hardwick et al, 1999). Similarly, mTOR has been identified as a direct target for the action of protein kinase B and of having a key role in insulin signalling (Shepherd et al., 1998; Nave et al, 1999). Mammalian TOR has also been implicated in the polarization of the actin cytoskeleton and the regulation of translational initiation (Alarcon et at, 1999). Phophatidylinositol 3-kinases, such as mTOR, are functional in several aspects of the pathogenesis of tumours such as cell-cycle progression, adhesion, cell survival and angiogenesis (Roymans and Slegers, 2001).
Most immunophilins do not appear to be directly involved in immunosuppressive activities and relatively little is known concerning their natural ligands although candidates for natural ligands of the FKBPs termed FKBP-associated proteins (FAP) such as FAP48 and FAP1 have been reported. The specific interaction of FAPs with FKBPs during the formation of complexes was prevented by rapamycin in a dose-dependent manner (Chambraud et at, 1996; Kunz et al., 2000). Immunophilins appear to function in a wide range of cellular activities such as protein folding; assembly and trafficking of proteins; co-regulation of molecular complexes including heat shock proteins; steroid receptors; ion channels; cell-to-cell interactions and transcription and translation of genes (Galat 2000; Hamilton and Steiner 1998). All immunophilins possess the protein folding property of peptidyl-prolyl cis-trans isomerisation and several immunophilins are found located in the endoplasmic reticulum, a principal site of protein synthesis in the cell. In addition to FKBP12 (U.S. Pat. No. 5,109,112) other immunophilins include FKBP12.6 (U.S. Pat. No. 5,457,182), FKBP13 (Hendrickson et al., 1993; U.S. Pat. No. 5,498,597), FKBP25 (Hung and Schreiber, 1992; Jin et al, 1992), FKBP14.6 (U.S. Pat. No. 5,354,845), FKBP52 (U.S. Pat. No. 5,763,590), FKBP60 (Yem et al, 1992) and FKBP65 (Patterson et al., 2000).
The multitude of the FKBP's which are present in different cell types also underline the utility of isolating novel FKBP-ligand analogues with potentially changed binding and/or effector domains.
Pharmacokinetic studies of rapamycin and rapamycin analogues have demonstrated the need for the development of novel rapamycin compounds that may be more stable in solution, more resistant to metabolic attack and have improved bio-availability. Modification using chemically available positions on the molecule has been addressed, however, this approach has limited utility as the sites available for chemical modification are limited and there is less ability to selectively modify a particular position. Biological approaches to producing novel rapamycin analogues have been less successful due to the difficulties encountered in working with the organism (Lomovskaya et al., 1997; Kieser et al., 2000) despite the availability of the sequence of the biosynthetic gene cluster of rapamycin from S. hygroscopicus (Schwecke et al., 1995).
A range of synthesised rapamycin analogues using the chemically available sites of the molecule has been reported. The description of the following compounds was adapted to the numbering system of the rapamycin molecule described in FIG. 1. Chemically available sites on the molecule for derivatisation or replacement include C40 and C28 hydroxyl groups (e.g. U.S. Pat. No. 5,665,772; U.S. Pat. No. 5,362,718), C39 and C16 methoxy groups (e.g. WO96/41807; U.S. Pat. No. 5,728,710), C32, C26 and C9 keto groups (e.g. U.S. Pat. No. 5,378,836; U.S. Pat. No. 5,138,051; U.S. Pat. No. 5,665,772). Hydrogenation at C17, C19 and/or C21, targeting the triene, resulted in retention of antifungal activity but loss of immunosuppression (e.g. U.S. Pat. No. 5,391,730; U.S. Pat. No. 5,023,262). Significant improvements in the stability of the molecule (e.g. formation of oximes at C32, C40 and/or C28, U.S. Pat. No. 5,563,145, U.S. Pat. No. 5,446,048), resistance to metabolic attack (e.g. U.S. Pat. No. 5,912,253), bioavailability (e.g. U.S. Pat. No. 5,221,670; U.S. Pat. No. 5,955,457; WO98/04279) and the production of prodrugs (e.g. U.S. Pat. No. 6,015,815; U.S. Pat. No. 5,432,183) have been achieved through derivatisation. However, chemical modification requires significant quantities of rapamycin template and, as a base and acid labile compound, it is difficult to work with. Where chemical derivatisation can be group selective, it is often difficult to be site selective. Consequently, chemical modification invariably requires multiple protective and deprotecive steps and produces mixed products in variable yields.
The isolation of rapamycin analogues using biological methods such as biotransformation and phage-based genetic modification has also been described. Isolation of minor metabolites from both mutant strains and rapamycin producing strains has provided small quantities of a number of rapamycin analogues. These strains are often low yielding and produce mixtures of rapamycin analogues. The isolation of 27-O-desmethylrapamycin and 27-desmethoxyrapamycin was reported from the culture supernatant of S. hygroscopicus NCIMB 40319 (Box et al., 1995). The antifungal activity of 27-O-desmethylrapamycin was lower than that of rapamycin but the inhibition of FKBP12 PPlase activity seemed to be increased. The inhibition of ConA-stimulated proliferation of murine splenic T cells and the inhibition of LPS-stimulated proliferation of murine splenic B cells was decreased when compared to rapamycin (Box et al., 1995). Similarly, antifungal activities of the rapamycin derivatives prolylrapamycin, 27-O-desmethylrapamycin and 27-desmethoxyrapamycin were lower than that of rapamycin (Wong et al., 1998). Rapamycin analogues (16-O-desmethylrapamycin, 27-O-desmethylrapamycin, 39-O-desmethylrapamycin, 16,27-O-bisdesmethylrapamycin, prolylrapamycin, 26-O-desmethylprolylrapamycin, 9-deoxorapamycin, 27-desmethoxyrapamycin, 27-desmethoxy-39-O-desmethylrapamycin, 9-deoxo-27-desmethoxyrapamycin, 28-dehydrorapamycin, 9-deoxo-27-desmethoxy-39-O-desmethylrapamycin) were also isolated from Actinoplanes sp N902-109 after the addition of cytochrome P450 inhibitors and/or precursor feeding to the culture or after biotransformation of isolated rapamycin (Nishida et al, 1995). The use of such inhibitors, however, only allows the targeting of a particular enzyme function and is not site selective. Rational production of a single selected analogue is not possible via this method. The resulting production of mixtures of rapamycin analogues rather than a single desired product also impacts yield. The mixed lymphocyte reaction (MLR) inhibitory activity of the compounds was assessed and little effect on the activity was detected after the loss of the methyl group at C27 or/and C16. In addition, 9-deoxorapamycin showed a more significant decrease in activity and the loss of the methoxy group at C27, the hydroxy group at C28 and the substitution of a pipecolinyl group for a prolyl group resulted in a reduction in potency (Nishida et at, 1995). Similarly, biotransformation of rapamycin and the isolation of 16,39-O-bisdesmethylrapamycin have been reported (WO 94/09010). The retention of inhibitory activity in cell proliferation assays with compounds modified in the cyclohexyl ring, e.g. 39-O-desmethylrapamycin and C40 modifications such as SDZ RAD, identify this region of the molecule as a target for the generation of novel rapamycin analogues. Novel rapamycin analogues were reported after feeding cyclohexanecarboxylic acid, cycloheptanecarboxylic acid, cyclohex-1-enecarboxylic acid, 3-methylcyclohexanecarboxylic acid, cyclohex-3enecarboxylic acid, 3-hydroxycyclohex-4-enecarboxylic acid and cyclohept-1-enecarboxylic acid to cultures of S. hygroscopicus thus demonstrating the flexibility in the loading module of the rapamycin polyketide synthase (P. A. S. Lowden, PhD dissertation, University of Cambridge, 1997). These novel rapamycin analogues were produced in competition with the natural starter, 4, 5-dihydroxycyclohex-1-enecarboxylic acid, resulting in reduced yields and mixed products.
The isolation of recombinant S. hygroscopicus strains producing various rapamycin analogues, using biological methods mediated by phage technology (Lomovskaya et al., 1997), has been reported. In the presence of added proline derivatives, a S. hygroscopicus rapL deletion mutant synthesized the novel rapamycin analogues prolylrapamycin, 4-hydroxyprolylrapamycin and 4-hydroxyprolyl-26-desmethoxy-rapamycin (Khaw et al., 1998). Similarly, the novel rapamycins 3-hydroxy-prolyl-rapamycin, 3-hydroxy-prolyl-26-desmethoxy-rapamycin, and trans-3aza-bicyclo[3,1,0]hexane-2-carboxylic acid rapamycin have been identified as described in WO98/54308. The activity of prolylrapamycin and 4-hydroxyprolyl-26-desmethoxy-rapamycin was assessed in proliferation assays and the inhibitory activity of the latter compound was significantly less than that of rapamycin (Khaw et at, 1998). The deletion of five contiguous genes, rapQONML (responsible for post-polyketide modifications at C16, C27 and production of L-pipecolic acid) and their replacement with a neomycin resistance marker in S. hygroscopicus ATCC29253 using phage-based methology resulted in the production of 16-O-desmethyl-27-desmethoxyrapamycin when fed with pipecolic acid (Chung et al, 2001). No complementation of this deletion mutant has been demonstrated using this technology. Furthermore, the site-specific functionality of rapM and rapQ remains unclear; therefore, rational design of rapamycin analogues requiring methylation at C16-H or C27-OH has not been enabled. The phage-based methodology suffers from a number of drawbacks as described in more detail below. It offers a difficult and protracted process of obtaining engineered strains and has a reduced versatility in comparison to the methodology disclosed within this current patent.
Conventional approaches to manipulate rapamycin modifying genes using biological methods comprise the mutation or deletion of individual genes in the chromosome of a host strain or/and the insertion of individual genes as extra copies of homologous or heterologous genes either individually or as gene cassettes (WO01/79520, WO 03/048375). However, the isolation of novel rapamycin analogues using such biological methods has been limited due to the difficulties in transforming the rapamycin-producing organism S. hygroscopicus. It has been reported that the commonly used methods of transformation with plasmid DNA or conjugal transfer were unsuccessful with the rapamycin producing strain (Lomovskya et al., 1997, Schweke et al., 1995, Kieser et al., 2000). The current state of the art uses the methodology of Lomovskya et al. (1997), a work intensive phage based method that is severely limited by the size of the cloned DNA fragments transferred into S. hygroscopicus (Kieser et al., 2000). This technology is limited to the transfer of a maximum of 6.4 kb of cloned DNA. Thus, when complementing a deletion mutant using this technology the artisan is limited to the inclusion of ˜2 functional genes in addition to desired promoter, regions of homology and resistance marker. The genetic information for the rapamycin biosynthetic gene cluster has been available since 1995 (Schwecke et al., 1995), however, limited progress in this area has been made (Khaw et al., 1998; Chung et al., 2001; WO01/34816).